Date

06 Jan 2020

Attendees

Goals

Review current tasks

Discussion items

Item	Who	Notes
Outstanding tasks	@Tim Whitlock	A new Validation release was done before Christmas. All completed tickets in the November 2019 Sprint were included. Those not completed were moved to the January Sprint the tickets for this sprint will be reviewed in the next validation meeting. It is likely that there will be some vascular access tickets added to account for the collection changes which will try to avoid the use of spreadsheets and is reducing the number of required fields. See /wiki/spaces/RR/pages/1310884383 for old process details and /wiki/spaces/RR/pages/1341030343 for the changes. The orphaned PV Memberships have still not been resolved. Also TNG-599 was also discussed which is an example of an orphaned membership and how it was created. When memberships and data were migrated to the new UKRDC dead patients where not added to the UKRDC. When a dead patient gets deleted from PV (as has happened as Shaun has been tidying the PV data) we end up with a membership close message from PV which creates an orphaned PV membership since there is and will never be any data for this patient it will remain an orphan. George Swinnerton will run a check now Shaun has tidied PV to ensure all patients are present on the UKRDC he will also sort the memberships this week. Error rendering macro 'jira' : Unable to locate Jira server for this macro. It may be due to Application Link configuration. is now deployed to live however in the process the automatic exports that run on Friday mornings broke. The exporter code was broken due to a typo that should have been picked up by lint and pre-deployment checks. james.griffin (Unlicensed)will look at how this might be checked/improved on Bamboo for future. In addition the custom SQL scripts were broken due to a column-name change that was part of the deploy. This means as part of a deployment impacts to automatic processes need to be checked on staging and live. This needs to be documented. George Swinnerton still hasn't responded to NHSBTs request about using gpg to encrypt files. The deployment of the updated SQLAlchemey code that corrects issues with jtrace delete relationships hasnot been deployed to staging. These changes affect the services module however the impact of the code changes seems to be restricted to delete functionality and not have any impact on the UI. This will be done by the end of the week. Andrew Atterton has now restored the deleted radar patient. After some issues with backing up radar live in order to test locally. This was a mixture of pgdump version mismatches and database size. The majority of the data base is size is caused by the logs table which is 80GB. This is due to the log everything approach. Previously Rapolas has purged redundant information and saved the purged data to an archive file in case it is required (it hasn't yet been needed) Andrew Atterton will look at this and see if he identify what to purge and we can set up a regular purge to prevent the log table growing too big. We also discussed a two stage approach to the logging where we have a archive logs database where most useful log data is made readily available and the less important data is archived in some way. This would ensure the main RaDaR database remains small but allow the historic logs to grow in a searchable way. james.griffin (Unlicensed) will look at what might be required to do this and how we would integrate it etc. Once the data has been successfully purged we can look at syncing radar live with staging as it hasn't been done recently. George Swinnerton has produced an initial prevalent count query from the renalreg database and passed it to David Pitcher for review. He has made some suggestions and also asked James Medcalf who hasn't come back with anything yet. George Swinnerton will follow this up. NHSBT have sent through a new file however we are not sure what the CAG status is. Before Christmas the sharing of data was on hold until NHSBT provided sufficient assurances to satisfy CAG on their security etc. Andrew will check with Retha and Tom. Barts wish to concentrate on providing a RDA feed suitable for PV and haven't migrated the extra data required for a full RDA feed for the first couple of quarters. Despite sending them the logins for sending test data through they haven't yet. As Gorge has pointed out we also need a sample Q3 full feed to check if there is any missing data etc even if they are planning to use the old system for the previous two quarters as it is possible some data hasn't been migrated etc. Kings still haven't configured the live feed yet and George is concerned that they may have not understood the need to send all 2018 patients including those who died in the year.
james.griffin (Unlicensed)	Workitem Issue	james.griffin (Unlicensed)raised an issue from Shaun about a workitem merge on the UKRDC. the incoming record has a PID of 9999999 this indicates it was not allocated a pid and so cannot be merged. Instead the workitem indicates the incoming record was matched and has a changing MRN but the demographics differed by too much to allow automatic matching/MRN change and need checking. In this case it appears the MRN is oddly formatted so could be a unit issue and the surname has changed. This type of workitem is not a candidate for merging but typically indicates one of two things the incoming file is wrong and the unit needs to correct their record or (more often) they have made several corrections to the patient record demographics and the MRN has changed. an MRN change assumes the demographics don't changed or raises a workitem.
George Swinnerton	Vascular Access	George has done the vascular access matching. Shaun has identified some changes required to improve it for next time. George will sort these now so they are ready for next time.
George Swinnerton	RaDaR Performance	The reason for the performance issue on RaDaR was unclear however due to the way radar handles the patient object when showing a list of patients it appears to do a lot of lookup queries and has some calculated fields. The sqlalchemy needs reviewing to see if pre-emptive lookups can changed to such that not all lookups are done initially but only if required etc.
james.griffin (Unlicensed)	AKI Data	The issue with the wrong CCG code being imported was impart caused by the lack of documentation which needs updating. james.griffin (Unlicensed) will look at this.
George Swinnerton	INS App	Nadim has asked about the ins app changes. These require some changes from Nick to the schema to support the information that will be returned. George Swinnerton will see if the changes can be done as a separate build before the dialysis changes to get the updates sooner. There is a meeting next week about the app.
George Swinnerton	Alpha Study	The alpha study is being resurrected in some way and they are asking for the data off the servers to see what they have before progressing. Alpha Study

Action items

George Swinnertonto sort orphaned PV memberships 13 Jan 2020
George Swinnerton to check all patients on PV are present on UKRDC since in theory they are all alive and should be receiving data13 Jan 2020
james.griffin (Unlicensed) to look at how bamboo could test the exporte as part of the deployment process/integration tests before deployment ensuring the issues as seen with typos are not deployed in future 13 Jan 2020
Andrew Atterton to look at purging the RaDaR logs following a similar pattern to that which Rapolas did 13 Jan 2020
james.griffin (Unlicensed) to consider options for processing the logs to separate archive log database (for discussion) designed for easy searching of important information which would hold older historic data to ensure the "live" logs remain manageable and the RaDaR backups are not dominated by the log table 13 Jan 2020
Andrew Atterton to check on status of NHSBT with CAG and whether the file can be processed. If yes to process the NHSBT file 13 Jan 2020
George Swinnerton to make changes required to the Vascular Access matching code and process and document it 20 Jan 2020
james.griffin (Unlicensed) to review the AKI documentation and update it appropriately 20 Jan 2020
Kings feed update
Barts feed update

Browser not supported